The bottom-feed in beaten-down tech is the large-cap software that has already turned: ServiceNow, Workday and Zscaler bottomed in April, are back above their 200-day averages and are in their first pullback, while Cadence and Akamai are turning at theirs. Intuit, Adobe, Oracle, HubSpot and The Trade Desk are still knives — below their 200-day averages, down between 33% and 73% on the year, with no base under them.

The Read

Data as of the September 25, 2026 settled close.

The line that sorts beaten-down tech is the 200-day average, not the size of the drawdown. Of the 133 names on the desk's software, cloud, security, ad-tech, AI, semiconductor and AI-compute lists, 71 trade at least 20% below their 52-week highs. The ones worth waiting on have already done the first two legs of bottom → pop → flag → breakout. ServiceNow, Workday and Zscaler made their lows in April and now sit 16.3%, 19.8% and 15.3% above their 200-day averages. Their one-year returns are still negative (−26.2%, −22.0%, −32.7%), so these are long trends in repair, not intact ones — the stage where a flag forms and a trend-hold entry gets decided.

This week's pullback came from rates, not from the companies. The Federal Reserve raised rates on September 16 (CNBC). On September 23 five-year Treasury yields crossed 5% and most maturities hit near two-decade highs (Bloomberg), and the market began pricing another hike in October (CNBC). Long-duration software gave ground across the board: all four cloud-software funds and all eleven security stocks and funds the desk tracks fell between Wednesday's and Friday's closes. ServiceNow and Workday slipped under their 20-day averages on half and a third of normal volume — the shape of a healthy flag. Zscaler is the exception: it fell 10.06% on Friday on 1.77 times its average volume. The cause is not established: its news, pulled Friday, carries nothing on the drop; no event is dated to it; and the subject of its September 24 8-K is not in any source on file.

Five names pass both tests — a real business and evidence the selling has stopped:

• ServiceNow is the cleanest: four of five reversal signs, a pullback on light volume, and its 50-day average (about $125.79) still under the price. It reports October 28.
• Zscaler has pulled back into the $188–$198 breakout zone the desk's September 22 deep dive named as its entry — but on a heavy-volume drop, not the quiet retest that dive asked for. Let the zone hold first.
• Workday is flagging on light volume. Reuters reported Silver Lake take-private talks on August 13, so part of its tape is a deal bet.
• Cadence sits on its 200-day average and on its year's most-traded price after a 15.3% week whose cause is not established. It reports October 26.
• Akamai reclaimed its 200-day in the September 21–23 rally, before its Anthropic compute contract was disclosed after the September 24 close; Akamai's 8-K puts the commitment at about $11.6 billion over seven-year terms (Bloomberg's headline said $12 billion). It is a slow grower with shrinking quarterly operating income, so it earns a flag entry, not a chase.

Broadcom is the report's best business and its most frustrating chart. Fiscal third-quarter revenue was $29.59 billion against $15.95 billion a year earlier (SEC 10-Q), yet the stock sits 4.2% under its 200-day average, 28.7% off its high, and up only 5.0% on the year after +101.0% over two. That is a long uptrend under test, not a bottom-feed. The trigger is a reclaim of the 200-day (about $368); the line is the $345.49 support shelf.

Avoid the knives. Intuit guided fiscal 2027 revenue growth to 9–10%, down from about 14%, and sits at an RSI of 29.6, 31.6% under its 200-day. Adobe sold off after beating and raising in September. Oracle's growth is real but its funding is now the story: free cash flow was −$23.7 billion in fiscal 2026, and its force majeure notice on the New Mexico build (CNBC, September 24) put AI data-center debt under fresh scrutiny (Bloomberg, September 25). HubSpot's trend broke on its August 5 report, and The Trade Desk closed at a new 52-week low on Friday. CoreWeave, the neocloud laggard, is basing but has not turned, and its financing costs are rising.

Semiconductor pullbacks are not bottoms. Applied Materials, Lam Research and Western Digital are 34.4%, 28.1% and 42.9% off their highs, but each is above its 200-day and up 143%, 146% and 326% on the year: deep pullbacks in intact trends. The flag-and-breakout playbook applies to them; the bottom-feeding one does not.

Dated catalysts: Accenture reports October 1; Cadence and Amkor October 26; ServiceNow October 28; HubSpot November 4; The Trade Desk November 5; Akamai in early November (provider dates differ, November 5 or 10); CoreWeave November 11; then, on provider estimates, Intuit November 19, Zscaler November 24, Workday December 1, Adobe and Broadcom December 9 and Oracle December 10. Microsoft's October 27 report opens the hyperscaler spending cycle that Oracle, Broadcom and CoreWeave trade on. The macro date is the October rate decision the market is now pricing.

July graded in one line: the July 5 quality list was right on four of seven (Atlassian, Workday, Zscaler, Unity), wrong so far on HubSpot and UiPath, right to call The Trade Desk the line between bargain and trap, and right to reject Oracle as a cash machine.

How the Board Splits

Tape columns update with each settled close; the Phase and Note columns are as of September 25, 2026.

Excluded from scope: the power, nuclear and utility names (Constellation, Vistra, NextEra, Oklo, NuScale, TransAlta), solar and storage (First Solar, Enphase, SolarEdge, Sunrun, Eos), telecom (T-Mobile, BCE, SK Telecom), CRISPR Therapeutics, Intuitive Surgical and Tesla ride in on the broad AI list and are left out. Quantum computing and small AI software (IonQ, Rigetti, D-Wave, BigBear.ai, SoundHound, C3.ai) are vetoed on business quality: lottery-style, early-stage stories, not bottom-feeds.

The Candidates

Five names where the business is real and the selling has visibly stopped. Every entry below is a flag condition; none is "buy now."

ServiceNow ($NOW): the cleanest turn

What it is: enterprise workflow software for IT, HR and customer-service operations.

Why it was knocked down: a year of selling on the view that AI agents replace software seats — the narrative the desk's July 5 and July 8 reports described. It was 47% off its high on July 7.

Evidence the selling stopped: it bottomed in April, reclaimed its 200-day, and shows four of five reversal signs (selling stopped, a reclaimed and rising 50-day, a higher low, RSI through its ceiling, beating the S&P 500 over one and three months). It is up 37.9% over three months and 16.3% above its 200-day. The pullback now under way sits under the 20-day ($138.75) and above the 50-day ($125.79), at RSI 51.6, on Friday volume about half of normal. Long horizon: −26.2% over one year, −23.8% over two, 43.4 points behind the S&P 500 over the year — a trend in repair.

Quality check: revenue +22.2% over the twelve months to June 30, 2026; GAAP operating margin 13.7% and free cash flow $4.58 billion in fiscal 2025 (SEC filings). One flag: June-quarter GAAP operating income was $162 million against $358 million a year earlier (SEC 10-Q); the reason is not established here, and the October 28 report is where it gets tested.

Entry: after the pullback holds the 50-day (about $125.79 on September 25), a close back above the 20-day (about $138.75). A close under the 50-day ends the setup.

Zscaler ($ZS): the retest arrived the wrong way

What it is: inline cloud security — user and application traffic runs through Zscaler's data centers for inspection; its 10-K opens by calling it "the cybersecurity platform for the AI era" (per the desk's September 22 deep dive).

Why it was knocked down: it fell 66% from its November high to an April low as growth slowed, including a 31.5% one-day drop on May 27 after a report that cut the free-cash-flow margin outlook on higher capital spending (the desk's September 22 deep dive). Friday's 10.06% drop: cause not established.

Evidence the selling stopped: two higher lows in May and June, a base between about $159 and $191 from mid-August, and a breakout on September 14 (September 22 deep dive). Even after Friday it is 15.3% above its 200-day and up 46.0% over three months. Long horizon: −32.7% over one year, +12.4% over two, 49.9 points behind the S&P 500 over the year.

Quality check: fiscal 2026 (ended July 31) revenue $3.35 billion, +25.4% (SEC 10-K, cross-checked; about 20% without the Red Canary acquisition, per the company); GAAP operating margin −4.0%; operating cash flow $1.13 billion (SEC). Fiscal 2027 guidance is about 17% growth (company).

Entry: the pullback holds $188 and closes back above $198 — the September breakout zone — on lighter volume than Friday's. A close below $170, the September 14 breakout-day low, ends it.

Workday ($WDAY): a flag with a deal attached

What it is: HR and finance software for large employers.

Why it was knocked down: the same AI-seat selling; it was 43% off its high on July 7.

Evidence the selling stopped: it bottomed in April, then delivered the earnings proof — a first-quarter beat with a raised subscription guide (the July 8 report). It is 19.8% above its 200-day and up 52.5% over three months; the last week (−2.3%) and month (−0.7%) are a sideways flag on Friday volume about a third of normal. Long horizon: −22.0% over one year, −22.2% over two.

Quality check: revenue +13.3% over the twelve months to July 31, 2026; fiscal 2026 (ended January 31) GAAP operating margin 7.5%, operating cash flow $2.94 billion, free cash flow $2.78 billion (SEC). Second-quarter fiscal 2027 revenue was $2.65 billion against $2.35 billion a year earlier (SEC 10-Q).

Caveat: Reuters reported on August 13 that Silver Lake was in talks to buy Workday. A signed deal turns this from a trend setup into a deal spread.

Entry: a close back above the 20-day (about $193.30) with the 50-day (about $180.93) intact. Below the 50-day, stand aside.

Cadence ($CDNS): on the line

What it is: chip-design (electronic design automation) software.

Why it was knocked down: it is 21.7% off its high and −7.1% on the year — a derating without a filed break. TipRanks records a second-quarter EPS miss against consensus; the cause of the drawdown is not established beyond that.

Evidence the selling stopped: its 52-week low is 162 sessions old. Price now sits on the 200-day (+0.2%) and on $317.27, the year's most-traded price and its nearest support shelf. It rose 15.3% this week; the cause is not established (Friday's news holds only an aggregator's unsourced claim, no company event is dated to the move, and followed sources are silent). Its peer Synopsys bounced 10.6% in the same week off a 52-week low eight sessions earlier. Long horizon: +20.0% over two years.

Quality check: revenue +13.4% over the twelve months to March 31, 2026; fiscal 2025 GAAP operating margin 28.2% and free cash flow $1.59 billion (SEC).

Entry: a pullback that holds $317.27, then a close above $352.55, the top of the price band where about 70% of the past year's volume traded. A close under $317.27 ends the setup.

Akamai ($AKAM): turned on a contract

What it is: content delivery, security and cloud computing.

Why it was knocked down: it is 31.1% off its high but still +51.2% on the year and 33.9 points ahead of the S&P 500 — a deep pullback inside an intact trend. The cause of the pullback is not established.

Evidence the selling stopped: it reclaimed its 200-day in the September 21–23 rally, before the deal, and holds it (+2.5%). The Anthropic compute agreement came after the September 24 close: Akamai's 8-K puts it at about $11.6 billion over seven-year terms, with a warrant that vests as Anthropic commits more; the "up to 20b potential" on X reads like that vesting ladder, not a signed amount. It is up 9.0% on the week, and Friday's volume ran 8.9 times average.

Quality check: revenue +5.9% over the twelve months to June 30, 2026; fiscal 2025 GAAP operating margin 13.5% and free cash flow $1.01 billion before $311.7 million of capitalized software (SEC). Second-quarter 2026 GAAP operating income was $80.3 million against $151.5 million a year earlier (SEC 10-Q). A followed account says Akamai expects about $5.5 billion of capital spending to serve the contract — unverified.

Entry: a flag that holds the $108.28 shelf, then a close above $126.02, the top of the past year's high-volume band. A close under $108.28 means the turn failed. It reports in early November.

Quick Dives: The Laggards That Surprise

Five laggards that surprise: strong businesses still sliding, or leaders left behind while their peers turned.

Intuit ($INTU): a cash machine the market has stopped paying for

What it is: TurboTax, QuickBooks, Credit Karma and Mailchimp, serving about 93 million consumers, small businesses and accountants (10-K). Its stated strategy is "an AI-driven expert platform" built on "a virtual team of AI agents and AI-enabled human experts."

Cratering, and since August. Intuit is 60.8% off its high, 31.6% under its 200-day, −60.2% over one year and −56.6% over two, 77.4 points behind the S&P 500, at RSI 29.6. The July–August bounce took it to $367.00 on August 22 (the desk's profile log); it has since given all of it back and is trading where it was in June and July ($280.99 on June 16, $274.96 on July 11). The desk has rated the Intuit thesis broken for seven straight weeks.

Why it is down. The August 25 fiscal fourth-quarter report set fiscal 2027 revenue growth at 9–10%, and the stock fell 11% that day (Bloomberg). The September 17 investor day reaffirmed that guide (StockTitan). Coverage framed it as a growth reset against about 14% in fiscal 2026 and cited AI-disruption fear and cut long-term targets (ad-hoc-news, an aggregator). The September 16 rate hike added pressure on long-duration software.

The numbers. Revenue grew 13.9% in the twelve months to July 31, 2026 (SEC, cross-checked). Fiscal 2026 operating cash flow was $8.84 billion (cross-checked) and free cash flow $8.66 billion (SEC). Operating margin is not quoted: the filing and the data provider disagree.

The April tax-season quarter carries the year; the November report covers the smallest quarter.

What the filing says. The September 9 10-K warns that competitors may "successfully use and deploy new technologies such as AI that may reduce customer demand for our products or services," and names government-run free filing as a risk: "although the IRS free direct filing system was suspended, proponents of these programs continue to advocate."

Followed sources. One account listed Intuit among "distribution moat" names where "AI bookkeeping and filing land on an installed base" (@pepemoonboy, September 20) — a view, not evidence.

Verdict: too early, with value-trap risk. The business still grows and throws off cash, but the price has no base, and buying a name at RSI under 30 right after it lowered its outlook is the washout pattern that has lost. What changes it: weeks of basing above the $262.24 support shelf and a close over the 50-day (about $324.65), then a November 19 report (provider estimate) that holds the 9–10% guide. A 26% operating margin sometimes quoted for Intuit is one of the figures the filing and the data provider disagree on, so it is left out.

Broadcom ($AVGO): the business accelerates, the stock does not

What it is: semiconductors — custom AI accelerators and networking silicon — plus infrastructure software including VMware (10-K). Its top five end customers were about 40% of fiscal 2025 revenue (10-K).

Out of favor, not cratering, since late August. Broadcom is 28.7% off its high and 4.2% under its 200-day, which it lost in late August (the desk's September 23 profile). It is +5.0% over one year and +101.0% over two, 12.3 points behind the S&P 500 over the year; the desk still rates the thesis intact. Its peers turned without it: Marvell and Arm are 61.2% and 46.5% above their 200-days.

Why it is down. The September 2 report beat, but the current-quarter revenue forecast disappointed (CNBC). On September 14 the CEO said AI revenue targets had not changed amid Anthropic's push to slow AI development (CNBC), and repeated it on September 21 ("Our targets haven't changed," Yahoo Finance). Rates did the rest.

The numbers. Revenue grew 48.7% in the twelve months to August 2, 2026 (SEC, cross-checked). Fiscal 2025 GAAP operating margin was 39.9% (cross-checked) and free cash flow $26.9 billion (SEC).

What the filing says. The 10-K flags that customers "may reduce the amount of products … due to reduced capital expenditure spending" and names U.S.–China trade tension as a demand risk.

Followed sources. One account put "essentially 0% probability" on the frontier labs slowing down and pointed to supply-constrained Nvidia and Broadcom growth (@aleabitoreddit, September 13) — opinion.

Verdict: a bottom-feed candidate on a trigger, and the trigger has not fired. It looks like a leader left behind; the tape says a leader in a flag below its 200-day. The condition is a close above the 200-day (about $368.35) that holds for a week; a close below the $345.49 shelf ends it. The December 9 report (provider estimate) tests the AI revenue outlook.

Oracle ($ORCL): growth funded by debt, priced like it

What it is: a database and applications company that now rents AI computing capacity through Oracle Cloud Infrastructure (10-K).

Cratering, and the financing is why. Oracle is 57.5% off its high, 16.7% under its 200-day, −52.9% over one year and −17.6% over two, 70.2 points behind the S&P 500; it fell 7.1% this week. The desk still rates Oracle's demand thesis intact; the price trend is broken.

Why it is down. The September 10 quarter beat: revenue up 30% and cloud infrastructure up 121%, with debt at $125 billion (CNBC), and the stock "still loses ground" (MarketWatch). On September 24 Oracle sent a force majeure notice on its Project Jupiter data center in New Mexico and fell 3% (CNBC); local officials had delayed power permits (Semafor), and Morgan Stanley said the notice puts loan and lease documents under fresh scrutiny for everyone financing the AI buildout (Bloomberg, September 25). An earlier attributed report said the build is at least seven months late with about $18 billion of related loans near 90 cents on the dollar (X post, September 19, unverified).

The numbers. Revenue grew 21.6% in the twelve months to August 31, 2026 (SEC, cross-checked). Fiscal 2026 GAAP operating margin was 30.6% (cross-checked). Operating cash flow of $31.98 billion against capital spending of $55.66 billion left free cash flow at −$23.69 billion (SEC). In the August quarter, capital spending was $28.50 billion against operating cash flow of $23.10 billion (SEC 10-Q); that 10-Q also shows the $20 billion at-the-market stock program fully used, $11.36 billion of customer prepayments, and remaining performance obligations of $664 billion against $455 billion a year earlier.

What the filing says. The 10-K's risk list includes "risks associated with our outstanding and future indebtedness" and the risk of not securing "data center capacity at affordable rates."

Verdict: a value trap until the financing leg stabilizes. What changes it: a base that reclaims the 200-day (about $164.54), and a December 10 report (provider estimate) where free cash flow is less negative without a new equity program.

Adobe ($ADBE): beat, raised, sold

What it is: Creative Cloud, Acrobat and marketing software, with its own Firefly image models (10-K).

Out of favor for two years, falling again since September 10. Adobe is 35.3% off its high, 10.4% under its 200-day, −33.5% over one year and −54.5% over two, 50.7 points behind the S&P 500, at RSI 36.6. It bounced from $221.54 on July 7 (the July 8 report) to a two-month high of $294.53 and has given most of it back since its September 10 report.

Why it is down. The third quarter set a revenue record of $6.76 billion, raised full-year targets and beat on adjusted EPS, yet the shares fell; management is trading some near-term recurring-revenue growth for free users, a bet it says pays off over 2027 (Yahoo Finance). "In this environment you can't just meet," one analyst told MarketWatch. A CEO transition is under way (ad-hoc-news).

The numbers. Fiscal 2025 GAAP operating margin was 36.6% and operating cash flow $10.03 billion (both cross-checked); free cash flow was $9.85 billion (SEC). Trailing revenue growth is not quoted: the filing and the provider disagree.

What the filing says. The 10-K names "AI or cloud-native companies" among competitors and expects competition "to continue to intensify."

Followed sources. One account called it a stock "sold on fears AI tools would replace their products entirely" (@CKCapitalxx, September 10) — a characterization, not evidence.

Verdict: too early. The quarters grow and the margin holds; the market is re-pricing durability, not the numbers. What changes it: a base and a close above the 200-day (about $262.68), ideally after the December 9 report (provider estimate) shows the free-user trade paying.

CoreWeave ($CRWV): the fastest grower pays the most for money

What it is: a GPU cloud built for AI training and inference (10-K). Microsoft was about 67% of 2025 revenue, and remaining performance obligations were $60.7 billion at December 31, 2025 (10-K).

Basing, not turned, since late summer. CoreWeave is 42.8% off its high, 5.0% under its 200-day, −30.9% on the year and 48.1 points behind the S&P 500; its 52-week low is 41 sessions old, and it sits on its support shelf ($85.64, also the past year's most-traded price). Its neocloud peer Nebius is 44.6% above its 200-day.

Why it is down. The cost of money. Banks marketed a roughly $1.1 billion junk bond for a Blue Owl–sponsored data center leased to CoreWeave (Bloomberg, September 21); an attributed report says a $2.6 billion CoreWeave loan priced at SOFR plus 550 basis points (X post relayed September 20, unverified); the company filed an offering prospectus on September 17 (SEC). Oracle's force majeure notice then put AI data-center debt under scrutiny (Bloomberg, September 25).

The numbers. Revenue grew 115% in the twelve months to June 30, 2026 (SEC, cross-checked). In 2025, revenue was $5.13 billion, GAAP operating margin −0.9% and net margin −22.8% (cross-checked); operating cash flow of $3.06 billion against capital spending of $10.31 billion left free cash flow at −$7.25 billion (SEC). In August the company raised its 2026 revenue guidance to $12.4–13.2 billion (CNBC).

What the filing says. Its risk factors lead with "our substantial indebtedness" and state the company "will require additional capital to fund our business."

Followed sources. One account moved its neocloud thesis from CoreWeave to Nebius, citing CoreWeave's high-interest debt (@aleabitoreddit, September 24) — opinion.

Verdict: too early. A real, fast-growing business carrying balance-sheet risk. The minimum for a trend-holder is a close above the 200-day (about $92.18); a close below the $85.64 shelf fails the base. It reports November 11. The financing strain is now on the record in Bloomberg's reporting, not only in social posts.

Still Falling

The names still making lower lows are software sold on AI substitution, with rates as the accelerant. The charge is framed the same way across the coverage: AI-disruption fear around Intuit's investor day (ad-hoc-news), a stock "sold on fears AI tools would replace their products entirely" (a followed account on Adobe), and a UiPath outlook read as limited acceleration (Ticker Report). That is a narrative, not a measured loss: each of Intuit, Adobe, HubSpot and The Trade Desk grew revenue in its latest reported year or quarter (SEC). The market is paying less for each dollar of it, and a rate-hiking Fed lowers the price of distant earnings further.

• Intuit, Adobe, Oracle: see Quick Dives.
• HubSpot broke on its August 5 report, when net customer additions missed and guidance implied slower growth (Investing.com's recap, per the desk's September 23 profile). Revenue still grew 21.1% in the twelve months to June 30, 2026 (SEC, cross-checked). It is 14.7% under its 200-day and −57.8% on the year; it reports November 4.
• The Trade Desk made a new 52-week low on Friday. Its August 6 release said the second quarter did not meet the company's own standards. Revenue grew 11.6% in the twelve months to June 30, 2026, and fiscal 2025 GAAP operating margin was 20.3% (SEC, cross-checked). It is 45.8% under its 200-day and −73.1% on the year; it reports November 5.
• UiPath fell hard after its early-September report as traders read the outlook as limited acceleration (Ticker Report, September 5; Timothy Sykes, September 8) and has since slipped under its 200-day, −25.7% on the month; its news has not been pulled since September 9.
• Ad tech (Criteo, Taboola, MNTN) and government IT (Booz Allen, Leidos) are also below their 200-day averages and still falling; no cause is claimed for them here.

Grading July

The desk's July 5 software report named seven "quality on sale" names and rejected Oracle; the July 8 report made three calls (buy MongoDB, watch Workday, pass on Okta) and named five more as "one flag away"; the July 18 note set cybersecurity entries and excluded Zscaler. Three-month change runs from late June, about a week before the July 5 data.

The July read was right on direction and loose on execution. The rotation into cloud software and security it called in early July happened. Where it lost value was the call mechanics: a watch with no entry price (Workday), a stop the first pullback took out before the move (MongoDB), and a timing pass with no reopen price (Okta). The calls below name a price for every leg; the stop trade-off remains — near stops can whipsaw, distant ones let a wrong call run.

Desk Call

Entries are flag conditions, not drawdown buys; stops sit at the nearest structure — a support shelf, a breakout-day low, the 50-day — rather than a distant 200-day. Moving-average levels are named and shown at their September 25 values; they move with the tape.

Sources

Primary sources:

• Rates: CNBC, Fed delivers its first hike in 3 years (2026-09-16); Bloomberg, US bond-market slide deepens (2026-09-23); MarketWatch, bond yields surge above 5% (2026-09-23); CNBC, market sees next Fed hike in October (2026-09-23)
• ServiceNow: 10-Q, quarter ended 2026-06-30; 10-K, FY2025
• Zscaler: 10-K, FY2026; Q4 FY2026 release; 8-K filed 2026-09-24
• Workday: Reuters, Silver Lake in talks to buy Workday (2026-08-13); 10-Q, quarter ended 2026-07-31; 10-K, FY2026
• Cadence: 10-Q, quarter ended 2026-06-30; 10-K, FY2025; TipRanks earnings page
• Akamai: Bloomberg, Akamai surges on Anthropic AI deal (2026-09-24); 8-K filed 2026-09-24; 10-Q, quarter ended 2026-06-30; 10-K, FY2025; @aleabitoreddit; @CKCapitalxx
• Intuit: Bloomberg, Intuit sinks 11% on weak outlook (2026-08-25); StockTitan, investor day reaffirms FY27 guidance (2026-09-17); ad-hoc-news, investor day resets growth outlook; 10-K, FY2026 (filed 2026-09-09); 10-Q, quarter ended 2026-04-30; @pepemoonboy
• Broadcom: CNBC, Q3 earnings report (2026-09-02); CNBC, CEO on Anthropic's slowdown push (2026-09-14); Yahoo Finance, "Our targets haven't changed" (2026-09-21); 10-Q, quarter ended 2026-08-02; 10-K, FY2025; @aleabitoreddit
• Oracle: CNBC, 30% revenue growth as debt hits $125 billion (2026-09-11); MarketWatch, Oracle's stock still loses ground (2026-09-11); CNBC, force majeure notice (2026-09-24); Bloomberg, notice to weigh on data-center debt (2026-09-25); Semafor (2026-09-25); X post, Project Jupiter delay (2026-09-19); 10-Q, quarter ended 2026-08-31; 10-K, FY2026
• Adobe: Yahoo Finance, Q3 FY2026 earnings; MarketWatch, earnings leave Wall Street wanting more (2026-09-10); ad-hoc-news, CEO transition; Trefis (2026-09-24); 10-Q, quarter ended 2026-08-28; 10-K, FY2025; @CKCapitalxx
• CoreWeave: CNBC, Q2 2026 report (2026-08-11); Bloomberg, CoreWeave-tied junk bond (2026-09-21); X post, AI-infrastructure loan terms; 424B5 prospectus (2026-09-17); 10-Q, quarter ended 2026-06-30; 10-K, FY2025; @aleabitoreddit
• Still falling: HubSpot 10-Q, quarter ended 2026-06-30; The Trade Desk Q2 2026 release; The Trade Desk 10-Q; Ticker Report on UiPath (2026-09-05); Timothy Sykes on UiPath (2026-09-08)

Internal receipts:

• Tape (every live cell, and prose figures as of 2026-09-25): summaries{cloud-saas,cloud-etf-holdings,cybersec,web-cloud,adtech,ai-scan,semis,ai-infra,neoclouds,government-compute,tech-insider-buys,bargain-bin,insider-universe,cpu-shortage-supply-chain,lazarus}.json`; phase labels from the September 25 bottom screen over the same lists.
• Reconciled fundamentals (SEC XBRL vs Massive; agreement recorded per figure): the September 25 fundamentals pull; quarterly series from stocks{INTU,AVGO,ORCL,ADBE,CRWV,NOW,WDAY,AKAM}/edgar/key-financials.json`.
• 10-K text: stocks{INTU,AVGO,ORCL,ADBE,CRWV}/massive/10k-{business,risk_factors}.json`.
• News receipts (all fetched 2026-09-25, ok): stocks{INTU,AVGO,ORCL,ADBE,CRWV,AKAM,AMKR,NOW,ZS,WDAY,CDNS,ACN,TTD,HUBS,APP}/news/latest.json`; PATH fetched 2026-09-09; ASAN 2026-09-17; SAP 2026-06-21 (stale).
• Earnings dates: 2026-10-01-acn-earnings, 2026-10-26-cdns-earnings.md, 2026-10-26-amkr-earnings.md, 2026-10-28-now-earnings.md, 2026-11-04-hubs-earnings.md, 2026-11-05-ttd-earnings.md, 2026-11-05-akam-earnings.md, 2026-11-11-crwv-earnings.md, 2026-11-19-intu-earnings.md, 2026-11-24-zs-earnings.md, 2026-12-01-wday-earnings.md, 2026-12-09-adbe-earnings.md, 2026-12-09-avgo-earnings.md, 2026-12-10-orcl-earnings.md.
• Captures: 2026-09-24-cnbc-oracle-sends-force-majeure-notice-about-data-center-project--108367403, 2026-09-25-bloomberg-oracle-notice-to-weigh-on-data-center-debt-morgan-stanley-sa-tlxgnit96osn00.md, 2026-09-25-semafor-oracle-force-majeure-on-us-data-center-showcases-ai-infrastr-astructure-risks.md, 2026-09-19-thebuildout-oracle-project-jupiter-data-center-build-faces-seven-month-d-6a86ff15.md, 2026-09-11-cnbc-oracle-posts-30-revenue-growth-fueled-by-ai-cloud-demand-as--108361644.md, 2026-09-11-marketwatch-oracle-shook-off-fears-about-ai-spending-but-its-stock-still--mktw-0005227559.md, 2026-09-14-cnbc-broadcom-ceo-addresses-anthropic-s-slowdown-push-says-ai-rev-108362673.md, 2026-09-21-yahoofinance-our-targets-haven-t-changed-broadcom-s-ceo-pushes-back-on-th-m-112507800-html.md, 2026-09-10-marketwatch-adobe-x2019-s-latest-earnings-leave-wall-street-wanting-more--mktw-0005226795.md, 2026-09-21-bloomberg-coreweave-tied-data-center-project-kicks-off-junk-bond-offer-tlqcynkjh6vc00.md, 2026-09-20-thebuildout-ai-infrastructure-lenders-demand-tougher-terms-as-financing--a28f7ad4.md, 2026-09-24-bloomberg-akamai-surges-on-12-billion-anthropic-ai-computing-deal-clos-tlvxnyrkv2th00.md, 2026-08-25-bloomberg-intuit-sinks-11-on-weak-outlook-as-nvidia-earnings-loom-clos-tkcdw4r24u8c00.md, 2026-09-16-cnbc-fed-delivers-its-first-hike-in-3-years-plus-what-s-moving-st-108363844.md, 2026-09-23-bloomberg-us-bond-market-slide-deepens-pushing-yields-to-two-decade-hi-tltm9rkjh6v700.md, 2026-09-23-marketwatch-bond-yields-surge-above-5-as-wall-street-fears-more-fed-rate--mktw-0005249446.md, 2026-09-23-cnbc-market-sees-next-fed-hike-in-october-following-barr-comments-108366836.md; tweets 2026-09-24-aleabitoreddit-akam-signs-11-6b-compute-deal-with-anthropic-for-48001701, 2026-09-24-CKCapitalxx-wow-akam-just-signed-a-7-year-11-6b-deal-with-an-32798924.md, 2026-09-13-aleabitoreddit-i-assign-essentially-0-probability-that-anthropi-04122410.md, 2026-09-10-CKCapitalxx-adbe-q3-fy26-revenue-6-76b-13-yoy-record-gaap-ep-13050499.md, 2026-09-24-aleabitoreddit-wow-nbis-relative-outperformance-is-wild-back-in-63133062.md, 2026-09-20-pepemoonboy-the-market-has-changed-in-my-eyes-the-biggest-ad-24422418.md; filings 2026-09-17-crwv-424b5-0001628280-26-062362, 2026-09-24-zs-8-k-0001713683-26-000193.md, 2026-09-24-akam-8-k-0001193125-26-401048.md.
• Desk context: 2026-09-22-zs-deep-dive; 2026-09-23-{INTU,AVGO,ORCL,HUBS}.md; Oracle 10-Q detail via the ai-capex-digestion campaign evidence (2026-09-11 Oracle profile); 2026-09-25-{cloud-etfs,cybersec,ai-scan,bargain-bin,tech-insider-buys,washout-scan}.md.
• July grading: 2026-07-05-software-wipeout-quality-on-sale, 2026-07-08-cloud-saas-rotation-watch, 2026-07-18-cybersec-entry-conditions-written-triggers; call outcomes (MDB stop and return from call, WDAY return from call, 30-day grades) from desk-calls (2026-09-25 build).